Booking a Procedure Around a Decision You Do Not Control

Booking a procedure in gastroenterology means committing an expensive resource to a date before you know whether it will be paid for. The room, the anesthesia coverage, and the physician’s block are all yours to schedule. The answer about coverage arrives on somebody else’s clock, in somebody else’s queue, on a timeline nobody at your practice can shorten.

Practices solve this two ways and both of them cost money. Some book immediately and absorb the cancellations when the answer comes back wrong or late. Others hold everything until the answer arrives and watch the procedure schedule run at partial capacity while patients wait.

Neither is a scheduling policy. They are two different ways of not having one.

The third option is to schedule against the shape of the decision rather than against its outcome. You cannot control when a determination comes back, but you can know which requests are pending, roughly how long that class of answer takes, and how long the answer stays good once it arrives. That is enough to build a rule.

The decision has a known shape even when the answer is unknown

Review programs increasingly publish their own timing, which turns an open-ended wait into a planning input. CMS is running the WISeR (Wasteful and Inappropriate Service Reduction) Model for six performance years, from January 1, 2026 to December 31, 2031, in six states: New Jersey, Ohio, Oklahoma, Texas, Arizona, and Washington. CMS describes the target as waste, which it says contributes up to 25% of health care spending in the United States.

The part a scheduler can use is the operating detail rather than the policy. Determinations under the model are typically issued within a few days of receipt, with a faster path when a delay would put the patient at risk, and an approval stays valid for a fixed number of months rather than indefinitely.

A practice that knows those two numbers can schedule. The earliest defensible offer is far enough out that the answer will have arrived, and the latest defensible offer is inside the window the answer will cover.

Everything between those two dates is bookable. Everything outside them is a promise you cannot keep.

Two earliest dates, not one

The rule that works in practice is that a GI practice has two different earliest-offer dates and has to know which one applies before it speaks.

A visit needing no review gets booked same day or next business day. A visit that does gets its earliest offer set out far enough for the authorization team to submit and hear back. In practice that generalizes to plan type as much as to service, with some plans routinely slower than others, so the real rule is the appointment type and the plan together rather than either one alone.

This is the opposite of how most scheduling logic is written. Most systems compute the earliest opening and offer it, which is correct for an office visit and wrong for a procedure that carries a review requirement.

On athenaOne this is a readable decision rather than a guess. The appointment type carries the service, the coverage on file carries the plan, and the open slot list carries the dates. The automation reads all three before it offers anything, and it offers from the correct pool rather than from the front of the calendar.

The rule that only moves one direction

Here is the constraint that catches practices out. An appointment whose authorization has not come back can be pushed later. It cannot be pulled earlier.

That sounds obvious written down and it is invisible in the moment. A patient calls, a cancellation has opened a slot next week, the scheduler is trying to help, and the pending request has not been decided. Moving the patient into that slot creates a procedure that will happen before the answer exists.

Rescheduling also breaks the link between the appointment and any approval already granted. The new booking is not attached to the existing authorization, so a move that looks like a favor produces an unattached procedure at checkout.

So the automation enforces direction. Pending requests can be offered later dates and cannot be offered earlier ones. Already-approved procedures can move anywhere inside the validity window and nowhere outside it. When a patient asks for something the rule forbids, it does not argue and it does not silently refuse. It books what it can and hands the exception to a scheduler with the pending status, the current date, and the earliest legal alternative attached.

Procedure blocks have edges the calendar does not show

Coverage is not the only constraint on a GI procedure slot, and the others are the ones that make generic scheduling tools fail.

Procedure blocks have hard cutoffs driven by room prep and turnover rather than by clinic hours, so the last genuinely bookable procedure of the day lands well before the practice closes. A tool reading the open slot list does not know that and will happily offer the late one.

Slot duration is the second trap. Changing a slot to fit a longer case means freezing or removing the neighbouring slots to make room, which is a change to the template rather than a booking. Outbound offers should never be making that change on their own.

And the appointment type catalog moves. One practice retired an entire catalog of procedure types overnight and folded them into a single short follow-up type, which cannot hold a long case. Any rule written against type names has to survive that, which means it reads the catalog rather than remembering it.

The automation works inside the boundaries it is given. It offers what the type and the cutoff allow, it does not reshape the template, and when filling the gap would require changing the day’s structure it raises that to the person who owns the block.

Tell the patient which clock they are on

The hardest part of this is not the scheduling logic. It is that the patient hears a date and stops listening.

A GI patient who has been told they need a procedure wants it handled. What they get instead is a conditional date, and if nobody explains the condition they will assume it is firm and plan around it.

So say the condition once, in plain terms, at the moment of booking. The date is held, the coverage decision is pending, and here is when we expect to know. Then follow up when the answer lands rather than waiting for the patient to call and ask.

That follow-up is a logistics call and should stay one. It confirms the date, tells the patient what to bring and how to prepare according to the instructions the practice already issues, and hands anything that turns clinical straight to staff. It is not the place to explain a payer’s reasoning or to advise the patient about the procedure itself.

What a practice should be able to see, and what to ask a vendor

Practice leaders are already prioritizing this kind of access work. Asked where they would focus on patient access in 2026, no-shows was the largest single answer at 27%, with online scheduling at 24%, phone access at 22%, and wait times at 21%. A procedure that gets cancelled because its coverage answer arrived late shows up in three of those four.

The view worth building is small. Pending requests with a date already held, approved procedures whose date sits outside the window, and approvals running out with nothing booked against them.

When you evaluate anything that claims to schedule procedures on athenaOne, ask it to book a case that requires review and watch whether its earliest offer changes. Then ask for a reschedule that would move the case earlier than the pending decision and see whether it refuses.

That behavior needs the scheduling side and the coverage side read together. PGA works across 440+ of athenahealth’s roughly 800 endpoints, which is why the open slots, the appointment types, and the authorization status can inform one offer instead of three disconnected screens.

Key Takeaways

Schedule against the shape of the decision rather than its outcome, using how long the answer takes and how long it stays valid as the two boundaries.
Run two earliest-offer dates, one for visits needing no review and a later one for those that do, set by appointment type and plan together.
Enforce direction on pending cases, since an appointment awaiting a decision can be pushed later but never pulled earlier.
Re-attach the authorization every time a procedure moves, because rescheduling breaks the link and produces an unattached case at checkout.
Respect procedure block cutoffs driven by room prep and turnover, which land well before the practice closes, and never let an offer reshape the template.
Tell the patient at booking that the date is held and the decision is pending, then call them when the answer lands instead of waiting for them to ask.

You cannot make a payer decide faster, and you can stop letting an undecided request take a procedure slot with it. Set two earliest-offer dates, let pending cases move only later, re-check the pair on every move, and give the patient the condition out loud at the moment you give them the date.

Sources